To run graphene, you need to do three things: partition a graph, format and mount disks and distribute graph data on disks, run graphene. These three steps also constitute the entire Graphene project structure as below.
- converter: convert a text tuple list into row-column balanced 2d partition format of the graph.
- disk_management: format and mount disks to the server and further distribute the partitioned dataset into these disks.
- Graphene contains lib and test two source code. Please find the detailed specification in the sub folders.
Should you have any questions about this project, please contact us by [email protected].
[FAST '17] Graphene: Fine-Grained IO Management for Graph Computing[PDF] [Slides]